Spot on. I programmed for a group of "idea" guys for while. Many a time I was torn off a work-in-progress only to begin another work-in-progress. It was so bad, I was only able to finish one project of dozens.

The key isn't having the ideas, starting the ideas, or even almost completing ideas. It's all about finishing.

Since I left, I now have one working idea; it's an amazing difference. I hope to have a MVP out soon, and I can't wait.

A couple months ago I did some back-of-the-envelope calculations to see the relative benefits of focusing on one idea versus doing many at once, and in most scenarios, focusing on one idea has the highest probability of success.

See the analysis here: http://www.startuptrekking.com/2010/05/startup-analysis-is-s...

The problem with just one idea is that the success of that idea is also subject to external factors beyond your control (ie competition).

Thus diversifying with multiple ideas, similar to in a stock portfolio, allows these externalities to cancel themselves other out.